AWS announced that it has signed a definitive agreement to acquire DuckLabs, the Amsterdam-based company behind DuckDB, the open-source analytical database that runs in-process and executes SQL queries directly on Parquet, CSV, and JSON files. According to the article, DuckDB will remain open source under the oversight of its independent foundation and under the MIT license.
Over time, AWS plans to combine DuckDB’s speed for everyday queries with the scale of its enterprise services, such as Amazon S3, Amazon Redshift, and Amazon Athena. DuckDB can be used locally or directly with Amazon S3, making it suitable for queries that typically do not exceed one terabyte, a category that accounts for most real-world analytics usage, according to the report.
DuckDB was co-founded by Hannes Mühleisen and Mark Raasveldt, and the two founders will continue to lead the project’s technical direction. AWS also pointed to the possibility of integrating DuckDB with Amazon EMR, AWS Glue, and Amazon SageMaker, and said the engine’s nature could make it suitable for intelligent-agent applications that explore data and iteratively experiment with queries.
What Changes in Practice?
The announcement does not indicate any immediate change to DuckDB’s license or development model; instead, it confirms that it will remain open source under the independent foundation. The potential longer-term change is the connection of a fast engine for local or data-adjacent queries with AWS’s broad range of services, but the article does not specify particular products or timelines for delivering this integration.
For AWS users, the report covered several practical updates during the past week:
- Amazon ECS: It now automatically detects when container instances lose connectivity with the agent and displays a health event named AGENT_CONNECTIVITY. ECS on Fargate and Managed Instances drains tasks, launches replacements, and deregisters the affected instance, while EC2 users can connect the event to their own workflows. The feature is available at no additional cost in all AWS commercial Regions and the GovCloud (US) Regions.
- AWS Lambda: It has made public previews available for the Node.js 26 and Python 3.15 runtimes. The upcoming versions can be tested before general availability using the same identifiers that will later be used, but AWS warns that breaking changes may occur, so these environments are not yet suitable for production.
- AWS IoT Core: It added a native action for routing device data to InfluxDB, whether through Amazon Timestream-managed or self-hosted deployments, with support for aggregation on the device or server side without writing custom code or creating an intermediary service.
- Amazon GameLift Servers: It received automatic protection against Layer 3 and Layer 4 DDoS attacks, such as UDP reflection and SYN floods, using AWS Shield Standard and game-appropriate traffic shaping, at no additional cost.
- Amazon SageMaker HyperPod: It expanded Ray support with capabilities for observability, elastic training, and accelerated inference, in addition to automatic node recovery, stalled-job detection, multilevel checkpoints, and multilevel KV cache in Ray Serve. This requires HyperPod clusters managed by Amazon EKS.
Broader Context for Data and Agents
AWS also presented the Agentic Resource Discovery specification, an open specification under the Apache 2.0 license for describing and discovering agents, tools, and MCP servers distributed across clouds, on-premises infrastructure, and SaaS platforms. AWS contributed feedback to the specification but does not own it. It complements the AWS Agent Registry by enabling connections between multiple catalogs without moving them.
It is also now possible to run aws configure agent-toolkit from the AWS CLI to provide AI-powered coding tools, including Kiro, Claude Code, Codex, and Cursor, with up-to-date AWS knowledge and secure connectivity to thousands of AWS APIs through the AWS MCP Server.
Editorially, the report combines an important strategic deal with operational updates of varying impact. The clearest current value for users lies in improved reliability, a better development experience, and support for Ray workloads, while the effects of the DuckLabs acquisition on future AWS integrations remain an open question, with no evidence in the article of an immediate change to DuckDB or a timeline for integration.